EU AI Act
EU AI Act Article 15: Accuracy, robustness and cybersecurity
UpcomingApplies from 2 Dec 2027. Applies from 2 Aug 2028 for high-risk AI systems under Article 6(1) and Annex I.
High-risk AI systems shall be designed and developed in such a way that they achieve an appropriate level of accuracy, robustness, and cybersecurity, and that they perform consistently in those respects throughout their lifecycle.
Article 15, paragraph 1. Regulation (EU) 2024/1689, consolidated text of 27 July 2026. Read the full text on EUR-Lex
EUR-Lex text is reused under Commission Decision 2011/833/EU. Guidance is stored as an interpretation, not as a statute.
What evidence usually supports this
- Evaluation runs with metrics against agreed thresholds
- Robustness and security test results
Maps to the ACCURACY_ROBUSTNESS and CYBERSECURITY controls in the Assurance OS catalog.
How Assurance OS gates it
Evaluation runs are checked against thresholds you set. A metric below its threshold blocks the release.
This page shows an excerpt of the law and the evidence teams commonly keep for it. It is not legal advice and does not say whether your system complies. Read the full text and take legal advice for your case.
Gate Article 15 evidence in CI.
Free plan, no card. Assurance OS checks the evidence on every release and hands buyers a signed pack.